The main user for this package was Seafile.
In the meantime, Seafile switch to PyMySQL.
https://pypi.org/project/PyMySQL/
PyMySQL seems to be a replacement for python-mysqlclient, and while it may
not be fully compatible with the MySQL API, it may be that those APIs
wouldn't be used.
This change drops this package.
If there is enough usage/reason to bring it back, we can.
For python-mysqlclient, the tag-line/description is:
```
This is a fork of MySQLdb1.
This project adds Python 3 support and bug fixes. I hope this fork is
merged back to MySQLdb1 like distribute was merged back to setuptools.
```
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Bump host Cython version as well.
Add note near PKG_VERSION to remember to periodically update it.
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
This adds PKG_BUILD_PARALLEL:=0 to packages that depend on host Python
packages (HOST_PYTHON3_PACKAGE_BUILD_DEPENDS), because installing
packages with multiple concurrent pip processes can lead to errors or
unexpected results[1].
This also:
* Move HOST_PYTHON3_PACKAGE_BUILD_DEPENDS definitions to before
python3-package.mk is included
* Update Python folder readme to include PKG_BUILD_PARALLEL:=0
[1]: https://github.com/pypa/pip/issues/2361
Signed-off-by: Jeffery To <jeffery.to@gmail.com>
This evaluates the arguments passed to "go install" during Build/Compile
rather than when golang-package.mk is included.
This also changes build directory-related variables to be recursively
expanded, because PKG_BUILD_DIR depends on BUILD_VARIANT and so can
change during different runs of Build/Compile.
Signed-off-by: Jeffery To <jeffery.to@gmail.com>
Essentially, this is a re-spin from
https://github.com/openwrt/packages/pull/9797/
But a really trimmed down version.
Only the Py3 variant is added now, which makes the Makefile really small
now.
Cython is needed on the host, to cythonize some files.
The package needs targets with hard-float enabled. This is because on some
soft-float targets floating-point exception constants aren't defined.
We can define some dummy values, but that ends up being a bit too much
work.
So, for that, the package depends on HAS_FPU or KERNEL_MIPS_FPU_EMULATOR.
This way, numpy should be buildable on hard-float, or for MIPS on the FPU
emulator being compiled in.
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
This also removes a link to pypi.org; the previous sentence already
links to pypi.org, a second link in the same paragraph isn't necessary.
Signed-off-by: Jeffery To <jeffery.to@gmail.com>
The perl Configure file was matching GCC 10 against "1*" and treating it
as GCC 1, causing ABI breakage and segfaults.
Cherry-pick the upstream patch which fixes it to check against (e.g)
"1.*" instead, which will make it work for hundreds more GCC versions
to come.
https://github.com/Perl/perl5/commit/6bd6308fcea3541
"Adapt Configure to GCC version 10"
Also includes the previous commit just adding GCC 8 and 9 to one case:
https://github.com/Perl/perl5/commit/ae195500577d707
"Add gcc-8 and gcc-9 for FORTIFY_SOURCE"
Signed-off-by: Ken Wong <xinxijishuwyq@gmail.com>
The pycrypto package is not maintained; the last stable release was in
2013.
This also updates python3-cryptodome, from conflicting with
python3-crypto, to providing python3-crypto.
Signed-off-by: Jeffery To <jeffery.to@gmail.com>
This took a while to get running.
It turns out, the fix is mostly in libxslt.
lxml uses the xsltGetProfileInformation() function, which is disabled from
libxslt via --without-profiler.
This causes a runtime error, since it cannot find the symbol.
So, libxslt is also updated to re-enable the profiler.
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Alexandru Ardelean
Javier Marcet
Zbyněk Kocur
Jeffery To
Donald Hoskins
Jan Pavlinec
Rosen Penev
Daniel Golle
Michael Heimpold
Josef Schlehofer
Andy Walsh
Xiaobo Tian
Philip Prindeville
Peter Stadler
Ken Wong